[ 
https://issues.apache.org/jira/browse/FLINK-8360?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16373145#comment-16373145
 ] 

ASF GitHub Bot commented on FLINK-8360:
---------------------------------------

Github user tillrohrmann commented on a diff in the pull request:

    https://github.com/apache/flink/pull/5239#discussion_r170038982
  
    --- Diff: 
flink-runtime/src/test/java/org/apache/flink/runtime/state/TaskLocalStateStoreImplTest.java
 ---
    @@ -59,12 +59,18 @@ public void before() throws Exception {
                this.subtaskIdx = 0;
                this.rootDirs = new File[]{temporaryFolder.newFolder(), 
temporaryFolder.newFolder()};
     
    +           LocalRecoveryDirectoryProviderImpl directoryProvider =
    +                   new LocalRecoveryDirectoryProviderImpl(rootDirs, jobID, 
allocationID, jobVertexID, subtaskIdx);
    +
    +           LocalRecoveryConfig localRecoveryConfig =
    +                   new 
LocalRecoveryConfig(LocalRecoveryConfig.LocalRecoveryMode.DISABLED, 
directoryProvider);
    --- End diff --
    
    It's a bit strange that we have to pass in a `directoryProvider`, even 
though the `LocalRecoveryMode` is disabled. I think it would be better to have 
something like `LocalRecoveryConfig.disabled()` and 
`LocalRecoveryConfig.fileBased(directoryProvider)`.


> Implement task-local state recovery
> -----------------------------------
>
>                 Key: FLINK-8360
>                 URL: https://issues.apache.org/jira/browse/FLINK-8360
>             Project: Flink
>          Issue Type: New Feature
>          Components: State Backends, Checkpointing
>            Reporter: Stefan Richter
>            Assignee: Stefan Richter
>            Priority: Major
>             Fix For: 1.5.0
>
>
> This issue tracks the development of recovery from task-local state. The main 
> idea is to have a secondary, local copy of the checkpointed state, while 
> there is still a primary copy in DFS that we report to the checkpoint 
> coordinator.
> Recovery can attempt to restore from the secondary local copy, if available, 
> to save network bandwidth. This requires that the assignment from tasks to 
> slots is as sticky is possible.
> For starters, we will implement this feature for all managed keyed states and 
> can easily enhance it to all other state types (e.g. operator state) later.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to